Qiben Shan

Combating Overfitting of Adversarial Training Efficiently via Balanced Instance Adaptive Defense

Adversarial training, one of the most effective methods for enhancing neural network robustness, is typically formulated as a min-max game between an attacker and a defender. Despite its success, most adversarial training methods suffer from robust overfitting, leading to a significant gap in robustness between the tra...